Path: news1.icaen!news.uiowa.edu!uunet!in3.uu.net!206.229.87.25!news.sprintlink.net!news-peer.sprintlink.net!visi.com!news
From: nathan@visi.com (Nathan Mates)
Newsgroups: comp.sys.apple2.programmer
Subject: Re: Scrolling the SHR screen (on the GS)
Date: 12 Dec 1996 07:15:44 GMT
Organization: Vector Internet Services, Inc.
Lines: 64
Message-ID: <58obf0$mos@darla.visi.com>
References: <58o7uv$c3g@engr.orst.edu>
NNTP-Posting-Host: thumper.visi.com

In article <58o7uv$c3g@engr.orst.edu>, Evan Day <day@flop.ENGR.ORST.EDU> wrote:
>I'm in search of some fast assembly code to scroll the GS SHR screen
>up or down a variable number of scanlines.  I can just use a
>ScrollRect() call on my accelerated machine with fair results, but
>on stock 2.8MHz machines, the thing tears and looks ugly.

   [Tip: 70 columns, not 80+ make it much easier to quote]

   In general, DP blasting is the fastest you can do under normal
circumstances. It's not immediately the most flexible (you may want
to have a few similar versions depending on which of the 5 pages
a given 16-scanline block starts in, etc). It can also be interrupt
unfriendly, unless you reenable interrupts every 5 pages (I do this
in Fish/Toast, and background MODs aren't really affected). It does
require bank 01 shadowing of the SHR screen.

   Here's some sample source for this kind of stuff. It's a straight
slam-thru of one page, no up/down scrolling.

	php		; store this
	sei		; IMPORTANT! 
	tsc
	sta	EntryStack	;we mess with this
	tdc
	sta	EntryDP
	shortm
	sta >$00C005	;bank 1 stack & dp writing
	sta >$00c003	;bank 1 stack & dp reading
	longm
	lda	#$2000	;top of screen
	tcd		;destination page
	clc
	adc	#$FF	;pei starts at top of page
	tcs

	pei	$FE
	pei	$FC
	pei	$FA
* ... Fill in the pattern
	pei	$02
	pei	$00

	shortm		;and put the system back in a sane state
	sta >$00c004	;bank 0 writing
	sta >$00c002	;bank 0 reading
	longm
	lda	EntryStack
	tcs
	lda	EntryDP
	tcd		;restore initial values *before* 
	plp		;we turn interrupts back on

   One of the GS tech notes (#70) recommends a NOP after every 13
peis as speeding things up. [Or so the hardware geeks with logic
probes at Apple say; I've never bothered benchmarking various
versions] That whole technote is rather informative; it's available
at ftp://ftp.apple.com/dts/aii/tn/iigs/tn-iigs-070

Nathan Mates
--
<*> Nathan Mates http://www.visi.com/~nathan/      <*>
# What are the facts? Again and again and again-- what are the _facts_?
# Shun wishful thinking, avoid opinion, care not what the neighbors
# think-- what are the facts, and to how many decimal places?  -R.A. Heinlein